Pub Date : 2023-06-28DOI: 10.1109/JCSSE58229.2023.10201984
Pongsathorn Pichaiyuth, Puwa Termnuphan, Tuul Triyason, Olarn Rojanapornpun, S. Jaiyen
Investment predicated on price trends stands as one of the most prevalent and efficacious approaches, hinging on its capacity to accurately discern the price trajectory for each asset. Such a pursuit poses itself as one of the most formidable challenges within the realm of investments. In this study, the application of machine learning models is employed, while simultaneously comparing their prognostic capabilities to evaluate their performance in forecasting cryptocurrency price trends. Additionally, the normalization technique and the Shapley Additive exPlanations (SHAP) feature selection method are employed to effectively augment the aptitude for projecting cryptocurrency price trends. The prediction period encompasses the time span from January 1, 2014, to December 31, 2021. The experimental findings reveal that the Support Vector Machine (SVM) outperforms other models such as K-Nearest Neighbors (KNN), Random Forest (RFC), Naïve Bayes, and Long short-term memory (LSTM) when forecasting periods extend 7, 15, and 30 days beyond the present, respectively. However, when the forecast horizon is extended to 90 days, the LSTM model exhibits the most optimal performance.
{"title":"Price Trend Forecasting of Cryptocurrency Using Multiple Technical Indicators and SHAP","authors":"Pongsathorn Pichaiyuth, Puwa Termnuphan, Tuul Triyason, Olarn Rojanapornpun, S. Jaiyen","doi":"10.1109/JCSSE58229.2023.10201984","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10201984","url":null,"abstract":"Investment predicated on price trends stands as one of the most prevalent and efficacious approaches, hinging on its capacity to accurately discern the price trajectory for each asset. Such a pursuit poses itself as one of the most formidable challenges within the realm of investments. In this study, the application of machine learning models is employed, while simultaneously comparing their prognostic capabilities to evaluate their performance in forecasting cryptocurrency price trends. Additionally, the normalization technique and the Shapley Additive exPlanations (SHAP) feature selection method are employed to effectively augment the aptitude for projecting cryptocurrency price trends. The prediction period encompasses the time span from January 1, 2014, to December 31, 2021. The experimental findings reveal that the Support Vector Machine (SVM) outperforms other models such as K-Nearest Neighbors (KNN), Random Forest (RFC), Naïve Bayes, and Long short-term memory (LSTM) when forecasting periods extend 7, 15, and 30 days beyond the present, respectively. However, when the forecast horizon is extended to 90 days, the LSTM model exhibits the most optimal performance.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"159 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121985323","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-28DOI: 10.1109/JCSSE58229.2023.10201943
Kritsana Treechalong, T. Rakthanmanon, Kitsana Waiyamai
In recent years, Numerous stream clustering techniques have recently emerged. However these techniques do not utilize the valuable background knowledge provided by domain experts. Using knowledge in stream clustering offers several advantages, including enhanced accuracy and performance of the resulting clusters. The proposed method in this research is CCE-Stream, which incorporates background knowledge as constraints for incremental stream clustering. Instance-level constraints, including Must-Link and Cannot-Link constraints, are used to guide improved clustering behaviors in various operations. Constraint operators are introduced to handle evolving constraint characteristics. CCE-Stream introduces the concept of assigning colors to constraints and a new cluster representation called Color of Cluster (CoC). Experimental results on Covertype and Electricity datasets demonstrate increased F-measure and Purity.
{"title":"CCE-Stream: Semi-supervised Stream Clustering Using Color-based Constraints","authors":"Kritsana Treechalong, T. Rakthanmanon, Kitsana Waiyamai","doi":"10.1109/JCSSE58229.2023.10201943","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10201943","url":null,"abstract":"In recent years, Numerous stream clustering techniques have recently emerged. However these techniques do not utilize the valuable background knowledge provided by domain experts. Using knowledge in stream clustering offers several advantages, including enhanced accuracy and performance of the resulting clusters. The proposed method in this research is CCE-Stream, which incorporates background knowledge as constraints for incremental stream clustering. Instance-level constraints, including Must-Link and Cannot-Link constraints, are used to guide improved clustering behaviors in various operations. Constraint operators are introduced to handle evolving constraint characteristics. CCE-Stream introduces the concept of assigning colors to constraints and a new cluster representation called Color of Cluster (CoC). Experimental results on Covertype and Electricity datasets demonstrate increased F-measure and Purity.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129390292","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-28DOI: 10.1109/JCSSE58229.2023.10202003
Nuttapol Kamolkunasiri, P. Punyabukkana, E. Chuangsuwanich
Image classification models in actual applications may receive input outside the intended data distribution. For crucial applications such as clinical decision-making, it is critical that a model can recognize and describe such out-of-distribution (OOD) inputs. The objective of this study is to investigate the efficacy of several approaches for OOD identification in medical images. We examine three classes of OOD detection methods (Classification models, Confidence-based models, and Generative models) on the data of X-ray images. We found that simple classification methods and HealthyGAN perform t he best overall. However, HealthyGAN cannot generalize to unseen scenarios, while classification models still retain some performance advantage. We also investigate the type of images that might be harder to detect as out of scope. We found that image crop-outs while being easily identifiable by humans, are more challenging for the models to detect.
{"title":"A Comparative Study on Out of Scope Detection for Chest X-ray Images","authors":"Nuttapol Kamolkunasiri, P. Punyabukkana, E. Chuangsuwanich","doi":"10.1109/JCSSE58229.2023.10202003","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10202003","url":null,"abstract":"Image classification models in actual applications may receive input outside the intended data distribution. For crucial applications such as clinical decision-making, it is critical that a model can recognize and describe such out-of-distribution (OOD) inputs. The objective of this study is to investigate the efficacy of several approaches for OOD identification in medical images. We examine three classes of OOD detection methods (Classification models, Confidence-based models, and Generative models) on the data of X-ray images. We found that simple classification methods and HealthyGAN perform t he best overall. However, HealthyGAN cannot generalize to unseen scenarios, while classification models still retain some performance advantage. We also investigate the type of images that might be harder to detect as out of scope. We found that image crop-outs while being easily identifiable by humans, are more challenging for the models to detect.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130199490","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-28DOI: 10.1109/JCSSE58229.2023.10202023
S. Fugkeaw, Narongsak Moolkaew, Theerapat Wiwattanapornpanit, Thanyathon Saengsena, Pattavee Sanchol
Cloud service providers generally rely on firewall and IDS targeting on volume-based detection applied to all subscribers. However, there are various data processing requirements especially the different transaction volume and data format supported by different web applications or web services. The general volume-based rule is incapable to address the fine-grained DDoS attack detection for web applications required resilient detection based on the statistical policy of individual cloud client. This paper proposes a design and implementation of a Cloud-based DDoS Attack Detection and Prevention System called CloudGuard system that offers a more fine-grained detection based on the integration of our proposed volume-based analysis and statistical web profile-based approach. Specifically, we proposed a tree-based DDoS detection model to efficiently detect and give response to DDoS attacks happening in the cloud environment. Furthermore, our proposed system entails a preventive mechanism based on the preventive policy to handle the case detected. Finally, we conducted the experiments to substantiate that our proposed scheme is functionally correct and efficient in practice.
{"title":"A Resilient Cloud-based DDoS Attack Detection and Prevention System","authors":"S. Fugkeaw, Narongsak Moolkaew, Theerapat Wiwattanapornpanit, Thanyathon Saengsena, Pattavee Sanchol","doi":"10.1109/JCSSE58229.2023.10202023","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10202023","url":null,"abstract":"Cloud service providers generally rely on firewall and IDS targeting on volume-based detection applied to all subscribers. However, there are various data processing requirements especially the different transaction volume and data format supported by different web applications or web services. The general volume-based rule is incapable to address the fine-grained DDoS attack detection for web applications required resilient detection based on the statistical policy of individual cloud client. This paper proposes a design and implementation of a Cloud-based DDoS Attack Detection and Prevention System called CloudGuard system that offers a more fine-grained detection based on the integration of our proposed volume-based analysis and statistical web profile-based approach. Specifically, we proposed a tree-based DDoS detection model to efficiently detect and give response to DDoS attacks happening in the cloud environment. Furthermore, our proposed system entails a preventive mechanism based on the preventive policy to handle the case detected. Finally, we conducted the experiments to substantiate that our proposed scheme is functionally correct and efficient in practice.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130544329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-28DOI: 10.1109/JCSSE58229.2023.10202030
Seoulcha Ratmumad, W. Suntiamorntut
Cloud computing technology has become increasingly popular and has expanded the capabilities of closed-circuit television (CCTV) systems, especially with the emergence of applications that allow easy access to CCTV. Cloud computing and CCTV capabilities has led to the development of cloud-based video processing applications, including video processing for surveillance and security, which utilize artificial intelligence (AI) technology to detect events in surveillance cameras and convert them into user-friendly results. In order to improve the processing speed in surveillance platform and support future surveillance camera functionalities, this paper proposes an optimized cloud video processing pipeline that leverages Apache Kafka and Distributed File System (DFS) technologies. We conducted experiments by applying configuration parameters to the message-oriented middleware (MOM) task and compared our approach to existing research on our test machines. We used the Node.js framework to run data producers and consumers. The results demonstrate that our proposed concept can reduce latency and increase system throughput, with a throughput increase of 88.55% for SD resolution image and 190.75% for HD resolution image compared to existing research.
{"title":"Improvement of Message-Oriented Middleware (MOM) for the Surveillance Platform","authors":"Seoulcha Ratmumad, W. Suntiamorntut","doi":"10.1109/JCSSE58229.2023.10202030","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10202030","url":null,"abstract":"Cloud computing technology has become increasingly popular and has expanded the capabilities of closed-circuit television (CCTV) systems, especially with the emergence of applications that allow easy access to CCTV. Cloud computing and CCTV capabilities has led to the development of cloud-based video processing applications, including video processing for surveillance and security, which utilize artificial intelligence (AI) technology to detect events in surveillance cameras and convert them into user-friendly results. In order to improve the processing speed in surveillance platform and support future surveillance camera functionalities, this paper proposes an optimized cloud video processing pipeline that leverages Apache Kafka and Distributed File System (DFS) technologies. We conducted experiments by applying configuration parameters to the message-oriented middleware (MOM) task and compared our approach to existing research on our test machines. We used the Node.js framework to run data producers and consumers. The results demonstrate that our proposed concept can reduce latency and increase system throughput, with a throughput increase of 88.55% for SD resolution image and 190.75% for HD resolution image compared to existing research.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127850220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-08DOI: 10.1109/JCSSE58229.2023.10202035
Dong Huu Quoc Tran, Hoang-Anh Phan, Hieu Dang Van, Tang Van Duong, Tu Bui, Van Nguyen Thi Thanh
The sampling-based exploration strategy is the most effective for Unmanned Aerial Vehicles, Micro Aerial Vehicles, and other three-dimensional outdoor robots. Its objective is to send robots to cover new unexplored areas by planning an optimal destination and path using an optimal utility function. Sampling-based Frontier Detection and Next Best View theories are the most powerful among the existing strategies for autonomous exploring and mapping techniques. This study proposes an improved sampling-based method for indoor robot exploration. The base algorithm's sampling task is adjusted to generate samples until the Rapidly-exploring Random Trees (RRTs) endpoints become frontiers. These samples are then evaluated using the enhanced utility function. The information obtained from the environments is estimated using occupied and uncertain probability. The initial results indicate that our modified NBV approach achieves a significantly larger explored area while reducing distance and time on Gazebo platform-simulated environments. These findings show our proposed approach's promising autonomous exploration potential in 2D environments.
{"title":"An Enhanced Sampling-Based Method with Modified Next-Best View Strategy For 2D Autonomous Robot Exploration","authors":"Dong Huu Quoc Tran, Hoang-Anh Phan, Hieu Dang Van, Tang Van Duong, Tu Bui, Van Nguyen Thi Thanh","doi":"10.1109/JCSSE58229.2023.10202035","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10202035","url":null,"abstract":"The sampling-based exploration strategy is the most effective for Unmanned Aerial Vehicles, Micro Aerial Vehicles, and other three-dimensional outdoor robots. Its objective is to send robots to cover new unexplored areas by planning an optimal destination and path using an optimal utility function. Sampling-based Frontier Detection and Next Best View theories are the most powerful among the existing strategies for autonomous exploring and mapping techniques. This study proposes an improved sampling-based method for indoor robot exploration. The base algorithm's sampling task is adjusted to generate samples until the Rapidly-exploring Random Trees (RRTs) endpoints become frontiers. These samples are then evaluated using the enhanced utility function. The information obtained from the environments is estimated using occupied and uncertain probability. The initial results indicate that our modified NBV approach achieves a significantly larger explored area while reducing distance and time on Gazebo platform-simulated environments. These findings show our proposed approach's promising autonomous exploration potential in 2D environments.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128574883","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-08DOI: 10.1109/JCSSE58229.2023.10201983
Hoang-Anh Phan, Phuc Nguyen, Thu Hang Thi Khuat, Hieu Dang Van, Dong Huu Quoc Tran, Bao Lam Dang, T. Bui, Van Nguyen Thi Thanh, T. C. Duc
Due to its numerous successful applications in industries like robotics, and autonomous navigation, 3D mapping for indoor environments has undergone much research and development. The complexity of the environment, a real-time embedding issue, and positioning mistakes of the robot system affect the creation of an accurate 3D map of indoor. Our research proposes a method to improve the 3D map construction performance by fusing data from the Ultrasonic-based Indoor Positioning System (IPS), the Inertial Measurement System (IMU) of the Intel Realsense D435i camera, and the encoder of the robot's wheel using the extended Kalman filter (EKF) algorithm. A Real-time Image Based Mapping algorithm (RTAB-Map) is used to handle the combined data, with the processing frequency updated in time with the IPS device's position frequency. The results indicate that combining sensors data considerably increases the speed, accuracy, and quality of the 3D mapping process. Our research demonstrates the potential of the integration of diverse data sources may be a useful tool for producing high-standard 3D indoor maps.
由于其在机器人和自主导航等行业的众多成功应用，室内环境的3D地图已经经历了许多研究和发展。环境的复杂性、实时嵌入问题和机器人系统的定位错误影响了精确的室内三维地图的创建。本研究提出了一种利用扩展卡尔曼滤波(EKF)算法，将基于超声的室内定位系统(IPS)、英特尔Realsense D435i相机的惯性测量系统(IMU)和机器人车轮编码器的数据融合在一起，以提高三维地图构建性能的方法。采用RTAB-Map (Real-time Image Based Mapping algorithm)算法对组合数据进行处理，处理频率随IPS设备的位置频率及时更新。结果表明，结合传感器数据可以显著提高三维制图过程的速度、精度和质量。我们的研究表明，整合各种数据源的潜力可能是制作高标准3D室内地图的有用工具。
{"title":"A Sensor Fusion Approach for Improving Implementation Speed and Accuracy of RTAB-Map Algorithm Based Indoor 3D Mapping","authors":"Hoang-Anh Phan, Phuc Nguyen, Thu Hang Thi Khuat, Hieu Dang Van, Dong Huu Quoc Tran, Bao Lam Dang, T. Bui, Van Nguyen Thi Thanh, T. C. Duc","doi":"10.1109/JCSSE58229.2023.10201983","DOIUrl":"https://doi.org/10.1109/JCSSE58229.2023.10201983","url":null,"abstract":"Due to its numerous successful applications in industries like robotics, and autonomous navigation, 3D mapping for indoor environments has undergone much research and development. The complexity of the environment, a real-time embedding issue, and positioning mistakes of the robot system affect the creation of an accurate 3D map of indoor. Our research proposes a method to improve the 3D map construction performance by fusing data from the Ultrasonic-based Indoor Positioning System (IPS), the Inertial Measurement System (IMU) of the Intel Realsense D435i camera, and the encoder of the robot's wheel using the extended Kalman filter (EKF) algorithm. A Real-time Image Based Mapping algorithm (RTAB-Map) is used to handle the combined data, with the processing frequency updated in time with the IPS device's position frequency. The results indicate that combining sensors data considerably increases the speed, accuracy, and quality of the 3D mapping process. Our research demonstrates the potential of the integration of diverse data sources may be a useful tool for producing high-standard 3D indoor maps.","PeriodicalId":298838,"journal":{"name":"2023 20th International Joint Conference on Computer Science and Software Engineering (JCSSE)","volume":"260 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133924698","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：